First of all I have to say that on the whole I see the Google Farmer update as a positive change, a move towards a better quality of search, but I say “on the whole” as there is no doubt some collateral damage in the so-called ‘War on Spam’.
The power of the long tail, and Google’s drum banging of its “content is king” mantra spawned a whole era of netrepreneurs that saw the correlation between writing unique content tailored to search queries to earn traffic, then monetising that traffic either by affiliation or with Adsense (a Google product ironically).
So 1 page became 10, and 10 pages became 504,000 (mahalo.com).
Now it appears that era is over, and Google have found and deployed a new algorithmic method of analysing site/page quality. This is a revelation, and I have to say when you look at the top 25 list as covered by Sistrix I think they nailed it.
So what changed?
Over at SEOMoz, Rand and his research team have (to my mind at least) determined that it’s not a link analysis update, and in the post he speculates that:
- “sites whose pages had fewer and/or less intrusive blocks of advertisements on them tended to be in the winner bucket”
- “sites whose UI/design would likely be described as more modern, high quality, thoughtful and “attractive” were winners vs. the “ugly” sites that tended to be in the loser bucket.”
- “those that tended to attract “thin” contributions (think EzineArticles, Hubpages or Buzzle) lost”
I agree with all of those points, but as Rand continues in the post, it would be hard (and dangerous) to deploy that analysis algorithmically, especially the design factor.
I don’t think any one element was used to score and downgrade these sites, but I have been looking around for another voice that makes reference to something that struck me immediately:
Is Google considering Social Media Buzz when evaluating sites with lots of content?
Google appears to have invested heavily in understanding, listening to and interpreting social media buzz – with “Realtime” entries in Google search results fed from Twitter acting almost as an instant barometer of importance. To that end, I suggest that Google looks at brand mentions in the social sphere when evaluating sites.
If Google trusts the word of the masses so much in social circles, and has established that this metric would be very difficult to fake, doesn’t it stand to reason that it would use it as an informal quality score?
Looking very speculatively at brand mentions vs size of site we can compare some winners and losers of similar site size to create at least some food for thought.
M = Twitter Mentions (Note 1)
P = Pages in Google index (Note 2)
Z = ZWS Raw Buzz Score (P Divided by M multiplied by 1000)
M = 126
P = 113,000
Z = 1.1
M = 665
P = 63,500
Z = 10.5
M = 17,400
P = 20,000,000
Z = 0.9
M = 70,300
P = 21,900,00
Z = 3.2
Notes On Research techniques:
Note 1: I have used a custom date range to filter out the “Buzz” now appearing about these sites as a result of the update, so the mentions occurred between 01/01/10-31/12/10.
Query Used : “brandname” site:twitter.com -site:twitter.com/brandtwitter
Note 2: site:domain.com on Google.co.uk.
The Get Out Clause
These numbers draw speculative conclusions and of course there are lots of other social media sites to monitor, but I only include these basic workings and an opinion on the matter for the purposes of starting a discussion.
There are anomalies such as Hubpages, which have quite a bit of buzz (score 1.8) where Instructables have a score of 1.
But on the whole, in my brief workings the “Winner” sites had around 6 or 7 times more buzz than their “Loser” Counterparts.
On that basis, albeit without the in-depth research this question deserves I believe that the Google Farmer update has looked for:
- Social media buzz/brand mentions + links for bigger sites, just links from social media sites for smaller sites.
- Adsense code location on page (More intrusion, less credibility)
- Outbound links on page (Apply some “green light” ratio for number of affiliate/non-affiliate external links on a page)
- Internal links on the page to other deep, long page name articles
All of the factors above, if combined would wipe out almost all thin affiliates and content farms for the following reasons:
- Mr/Mrs A F Illiate would set up sites en mass, with a formulaic approach to adsense block placement (to try to improve CTR)
- He/she would not want to lose visitors by linking anywhere other than to their affiliate partners
- He/she would use software/wordpress plugin to link with anchor text rich links to other articles on their own site.
- As the sites are numerous, and require a “one-hit” add-5-pages-get-5-links-wait-to-rank approach, the webmaster would have no way of setting up a social media profile for the site, and as such would receive almost zero links from these places
The conclusions above are my own personal opinion, but certainly the sites that are/were penalised do fit quite nicely into all of the criteria in the bulleted list above.
Keen to hear any other opinions and findings around the topic of this update.
… but to quote Socrates:
True knowledge exists in knowing that you know nothing
Image Credit: Andrew Stawarz